Objective: To study the association between self-reported leisure time physical activity (LTPA) at baseline and change in body mass index (BMI).
Design: Prospective observational study with a 11-y follow-up period.
Setting: A total population-based health survey in one county was performed in 1984-1986 (HUNT 1) and repeated in 1995-1997 (HUNT 2).
Participants: In total, 21 685 men participated in both surveys. In the present study, we included only apparently healthy 20-69-y-old men participating in both surveys and who had a normal body weight (BMI 18.5-24.9 kg/m(2)) at baseline-leaving 8305 men for the analyses. In all, 6945 men answered all questions about LTPA and 6749 men had complete data in the multiple analyses.
Measurements and main results: At HUNT 1, the participants answered questions (self-reported) about the intensity, frequency and duration of LTPA. The association between change in BMI and LTPA was investigated in multiple linear regression analyses. Adjusted for smoking, education, age and BMI at baseline, the physical active cohort gained less weight than the inactive cohort. Low, moderate and high levels of LTPA showed a U-shaped effect adjusted for smoking, education, age and BMI at baseline. Adjusted for BMI and age at baseline, the high-intensity part of the physically active cohort gained less weight than the low-intensity group.
Conclusion: This study has demonstrated a moderate BMI effect of LTPA at the population level; however, even a high level of LTPA did not prevent weight gain during the 11-y follow-up period.